中文摘要：

为解决赤铁矿强磁选过程运行因工艺生产指标无法在线连续检测而难以实施闭环优化控制的问题，将过程知识、专家经验与过程数据相结合，设计了由基于案例推理的回路预设定模型、小波神经网络软测量模型以及专家规则补偿器构成的智能优化设定控制器．针对已有优化控制软件存在的标准化程度低、可扩展性差、资源无法复用等问题，设计并开发了基于．NET组件技术的集优化控制与运行监视为一体的磁选优化设定控制软件系统．该系统针对具有不同编程技能的研究人员，实现对动态链接库、Matlab脚本以及VBScript脚本3种算法形式的支持．在半实物仿真平台的实验研究表明，所开发的系统具有一定的鲁棒性，有效保证了最终产品质量．

英文摘要：

The closedloop optimal control of the technical indices for intensity magnetic separation process （IMSP） is difficult to realize since the fact that the indices cannot be measured continuously online. To solve this problem, an intelligent optimal setting controller, which consists of casebased reasoning （CBR） preset module, wavelet neural network （WNN） softsensor module and rulebased reasoning compensator, is designed with the integration of the process knowledge, expert experience and process data. In view of the drawback of the existing optimization control software system, such as lower degree of standardization, poor scalability and reusability of resources, a magnetic separation optimal setting control software system, which integrates optimization control and operation monitoring function, is designed and developed based on .Net component technology. This system supports three kinds of algorithm, i. e. , dynamic link library, Matlab script and VBScript script, according to differ ent people which have different program skills. The simulation experiments at hardwareinloop simula tion platform demonstrate that the system can guarantee product quality for IMSP under certain robust ness.